3 ms·

You can't build part of an engine in three months and then show it off on demo day. You can't build a database in three months and show it off on Demo Day eith

by pg 13y ago

You can't build part of an engine in three months and then show it off on demo day. You can't build a database in three months and show it off on Demo Day either, but we've funded many database companies. You don't need to have shipped by Demo Day. You just need to have achieved enough to convince investors to give you more money. What "enough" means depends on what you're building. Experienced investors know how much they can reasonably expect from each type of company in 3 + n months. Incidentally, we have battery and AI companies in the current batch.

temuze 13y ago
Someone should make a list of successful companies that needed a long time to develop! Knowing they exist is encouraging to people who think about committing that much time.
steve19 13y ago
Most databases I see on HN are a case of take algorithm X and apply it to problem Y through interfaces Z[], which is fantastic if you have problem Y and no good solution. Infinitely scalable nosql db running on raw SSDs is fantastic, for example, but not even close to designing a theoretical spaceship. Would or could ycombinator fund a company that could only tell you what to expect in 0.25 + n years where n>9 and Boeing, GD and ATK are all working on similar problems?
- pg 13y ago
  We'd fund anything that had big potential and effective founders. It might be harder to raise money after YC for something that was going to take a long time, but there are certainly some VCs who'd like an idea that was going to take a long time but had proportionately high expected value.
  - danelectro 13y ago
    I'm a new poster here, I started my company with all recycled equipment. Here is my stream of consciousness post about this. I could have thousands of words about it so this is intentionally brief. I am always trying to improve my abilities on this subject, and will discuss it more but this is what I've got right now. Breakthrough technology innovation is fundamentally different. This could be tough for teams which have been geared to a goal within sight. In natural science especially, it takes a research laboratory which has to be built in advance, and this means a much further stage of bug-free completion compared to release 1.0 of software (where traction may immediately begin). And then you get to start the pursuit of a/the breakthrough, over an unknown time-frame, only after which if you succeed, then you can begin to package and monetize it. When you "pivot", you usually need a new or additional laboratory, and be prepared to set R&D back to square one. Capital is needed for different things, in different ways, in different amounts, and at different times with differing terms than other high-profit or high-return operations. Compared to software engineering & marketing projects, these differences are so drastic, it's been interesting to me so I've been looking at it when I can. My interest in capital would be for natural science breakthroughs but to outsiders most available capital only appears on the "map" due to the search for the next facebook. I guess that's a concept where capital would be on a "map" and entities worth investing in would be good to have on their "radar". I wouldn't want to be shot down, I would want my technology to be guided in for a safe landing. Aviation fuel, and a wonderful takeoff. Some of these capital projects are so strong that tech capital is like Hollywood money & connections where the talent does not actually have to be that good any more. They can make a star out of anyone, and the system now tends to exclude the kinds of talent that made it great to begin with. It would be interesting to see what that type of west coast machine could do with one of my technologies, areas of expertise, or even simple innovative ideas. This has captured my imagination for a while now, but I guess I just have an over-active imagination. Ideas especially are not supposed to be worth anything on their own without action, but in natural science, if you are successful enough to have an actionable idea or breakthrough, and take action, it will usually be a longer-term committment than a software project, and before you have a chance to complete and/or leverage the development, most likely better ideas will come along which even if recognized as better will have to wait until later on for attention. If you are really advanced at things they don't teach in school, you will also have trouble getting your point across. Either way you end up with better ideas than anyone including yourself has had a chance to take action on or leverage in any way, so the best ideas lie dormant and unappreciated since they are just ideas and can not be worth anything compared to ones that have already been invested in. If your life's work is geared toward breakthrough innovation in natural science, you end up with a lab like Thos. Edison and not many people can relate to that any more. This is what has been on my mind for a couple hours now, since I first got the "RFS" in my email. I just subscribed to the Sam Altman blog not that long ago because I ran into some material of his which was inspiring. I'm not a social networker or a software engineer, and I'm not from a silicon valley industry, so I've been trying to get realistic ideas how these kinds of companies get off to such a good start with their capital. Now all of a sudden horizons here have broadened to include natural science, this is even more inspiring. I would expect some people would figure I'm too old for this ycombinator type stuff, but here I am now.
  - patrickk 13y ago
    Would you have to consider letting some companies that are executing these ideas reside outside of Silicon Valley for the YC program, due to geographical constrains? For example, let's say Mate Rimac, the founder of Rimac Automobili (http://en.wikipedia.org/wiki/Rimac_Concept_One http://en.wikipedia.org/wiki/Rimac_Concept_One) who is a Croatian guy scratching an itch by producing the type of bonkers performance EV that he wanted to become reality, came to yc and you liked him, but his team, lab, equipment etc was all back home in Croatia. What would you do? It seems like some of these ideas will be constrained by location.
- strlen 13y ago
  It's very easy to underestimate how difficult building a database from scratch (or even making large incremental changes to an existing one) is. Another way to think of it is to think of an core database engine as analogous to a part of a propulsion mechanism. You can "demo" the core to show that a technology has promise. In case of a DB, you'd use a micro-benchmark with a synthetic workload; in case of a propulsion mechanism, you could use a software simulation. The full RDBMS (transactions and locking, query optimization, buffer pool, etc...) takes far longer to build. Finally, the equivalent of a "spaceship" here would not be more than the RDBMS, but a full ecosystem of systems software, middleware, and applications (e.g., app servers, messaging, CRM, ERP, etc..)
x0x0 13y ago
hmm; if you're such an ai company, i'd love to hear from you so please post =P
malandrew 13y ago
What language(s) are the AI companies using?

by pg 13y ago

temuze 13y ago

Someone should make a list of successful companies that needed a long time to develop! Knowing they exist is encouraging to people who think about committing that much time.

steve19 13y ago

Most databases I see on HN are a case of take algorithm X and apply it to problem Y through interfaces Z[], which is fantastic if you have problem Y and no good solution. Infinitely scalable nosql db running on raw SSDs is fantastic, for example, but not even close to designing a theoretical spaceship. Would or could ycombinator fund a company that could only tell you what to expect in 0.25 + n years where n>9 and Boeing, GD and ATK are all working on similar problems?

pg 13y ago
We'd fund anything that had big potential and effective founders. It might be harder to raise money after YC for something that was going to take a long time, but there are certainly some VCs who'd like an idea that was going to take a long time but had proportionately high expected value.
- danelectro 13y ago
  I'm a new poster here, I started my company with all recycled equipment. Here is my stream of consciousness post about this. I could have thousands of words about it so this is intentionally brief. I am always trying to improve my abilities on this subject, and will discuss it more but this is what I've got right now. Breakthrough technology innovation is fundamentally different. This could be tough for teams which have been geared to a goal within sight. In natural science especially, it takes a research laboratory which has to be built in advance, and this means a much further stage of bug-free completion compared to release 1.0 of software (where traction may immediately begin). And then you get to start the pursuit of a/the breakthrough, over an unknown time-frame, only after which if you succeed, then you can begin to package and monetize it. When you "pivot", you usually need a new or additional laboratory, and be prepared to set R&D back to square one. Capital is needed for different things, in different ways, in different amounts, and at different times with differing terms than other high-profit or high-return operations. Compared to software engineering & marketing projects, these differences are so drastic, it's been interesting to me so I've been looking at it when I can. My interest in capital would be for natural science breakthroughs but to outsiders most available capital only appears on the "map" due to the search for the next facebook. I guess that's a concept where capital would be on a "map" and entities worth investing in would be good to have on their "radar". I wouldn't want to be shot down, I would want my technology to be guided in for a safe landing. Aviation fuel, and a wonderful takeoff. Some of these capital projects are so strong that tech capital is like Hollywood money & connections where the talent does not actually have to be that good any more. They can make a star out of anyone, and the system now tends to exclude the kinds of talent that made it great to begin with. It would be interesting to see what that type of west coast machine could do with one of my technologies, areas of expertise, or even simple innovative ideas. This has captured my imagination for a while now, but I guess I just have an over-active imagination. Ideas especially are not supposed to be worth anything on their own without action, but in natural science, if you are successful enough to have an actionable idea or breakthrough, and take action, it will usually be a longer-term committment than a software project, and before you have a chance to complete and/or leverage the development, most likely better ideas will come along which even if recognized as better will have to wait until later on for attention. If you are really advanced at things they don't teach in school, you will also have trouble getting your point across. Either way you end up with better ideas than anyone including yourself has had a chance to take action on or leverage in any way, so the best ideas lie dormant and unappreciated since they are just ideas and can not be worth anything compared to ones that have already been invested in. If your life's work is geared toward breakthrough innovation in natural science, you end up with a lab like Thos. Edison and not many people can relate to that any more. This is what has been on my mind for a couple hours now, since I first got the "RFS" in my email. I just subscribed to the Sam Altman blog not that long ago because I ran into some material of his which was inspiring. I'm not a social networker or a software engineer, and I'm not from a silicon valley industry, so I've been trying to get realistic ideas how these kinds of companies get off to such a good start with their capital. Now all of a sudden horizons here have broadened to include natural science, this is even more inspiring. I would expect some people would figure I'm too old for this ycombinator type stuff, but here I am now.
- patrickk 13y ago
  Would you have to consider letting some companies that are executing these ideas reside outside of Silicon Valley for the YC program, due to geographical constrains? For example, let's say Mate Rimac, the founder of Rimac Automobili (http://en.wikipedia.org/wiki/Rimac_Concept_One http://en.wikipedia.org/wiki/Rimac_Concept_One) who is a Croatian guy scratching an itch by producing the type of bonkers performance EV that he wanted to become reality, came to yc and you liked him, but his team, lab, equipment etc was all back home in Croatia. What would you do? It seems like some of these ideas will be constrained by location.
strlen 13y ago
It's very easy to underestimate how difficult building a database from scratch (or even making large incremental changes to an existing one) is. Another way to think of it is to think of an core database engine as analogous to a part of a propulsion mechanism. You can "demo" the core to show that a technology has promise. In case of a DB, you'd use a micro-benchmark with a synthetic workload; in case of a propulsion mechanism, you could use a software simulation. The full RDBMS (transactions and locking, query optimization, buffer pool, etc...) takes far longer to build. Finally, the equivalent of a "spaceship" here would not be more than the RDBMS, but a full ecosystem of systems software, middleware, and applications (e.g., app servers, messaging, CRM, ERP, etc..)

x0x0 13y ago

hmm; if you're such an ai company, i'd love to hear from you so please post =P

malandrew 13y ago

What language(s) are the AI companies using?